Are Visual Memory Deficits in Recent-Onset Psychosis Degenerative?

To The Editor: In the October 2019 issue of the Journal, Zanelli et al. (1) presented evidence for a decline in visual memory but not executive function at 10-year follow-up in schizophrenia and other psychoses. In partial agreement with this finding, our group has reported that deficits in cognitive control (a form of executive function) are stable in the first year of psychotic illness (2). Intrigued by the results of Zanelli et al., we conducted an additional examination of the 1-year time course of long-term visual memory deficits in psychosis using the Relational and Item-Specific Encoding task (for task details, see references 3 and 4). Using this task, we previously demonstrated disproportionate item recognition deficits after relational encoding relative to item encoding as well as reduced associative recognition in schizophrenia (4). In the present analysis, performance (d′) was analyzed at baseline and at 12-month follow-up in a cohort of healthy control subjects (N=29; mean age, 18.34 years, SD=3.55; 20 men, nine women; average days to follow-up, 345, SD=65) and patients with recent-onset (<2 years from the first episode) schizophrenia spectrum disorders or type I bipolar disorder with psychotic features (N=28; 23 patients with schizophrenia, and five patients with bipolar disorder; mean age, 18.82 years, SD=4.10; 18 men, 10 women; average days to follow-up, 335, SD=102). The main effect of group (control compared with psychosis), time, and the group-by-time interaction were analyzed by repeated measures analysis of variance. Stability between time points was analyzed by calculating intraclass correlation coefficients (ICCs) (with absolute agreement) for each group.

Results are presented in Table 1. Greater than 50% accuracy was observed for all subjects during all task conditions at both time points. Significant main effects of group were observed for item recognition following item encoding, item recognition following relational encoding, and associative recognition. No main effects of time or group-by-time interactions were observed. Stability was good (ICC=0.60–0.75) to excellent (ICC<0.75) for item recognition and fair (ICC=0.40–0.60) for associative recognition in individuals with psychotic disorders. Between-group effect sizes either shifted from large (>0.8) to medium (between 0.5 and 0.8) or remained in the medium range between baseline and follow-up.

These results suggest that deficits in long-term visual memory are stable and nondegenerative at 1-year follow-up in recent-onset psychosis. It should be noted that the present sample was approximately 10 years younger than the baseline first-episode sample in Zanelli et al. It is possible, therefore, that the accelerated decline in memory observed by Zanelli et al. was affected by an interaction of age with illness (i.e., perhaps individuals with psychosis in their 30s are more susceptible to decline than teenagers or those in their 20s). Furthermore, nonpsychosis-related effects (e.g., smoking, diet, or metabolic syndrome) may have contributed to degeneration in the study by Zanelli et al. As our analysis examined individuals tracked only for 1 year, longitudinal examinations with longer follow-up periods using the Relational and Item-Specific Encoding paradigm will be required to test these possibilities.